Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution Paper • 2605.15138 • Published 12 days ago • 6
Mean Mode Screaming: Mean--Variance Split Residuals for 1000-Layer Diffusion Transformers Paper • 2605.06169 • Published 19 days ago • 229
From Context to Skills: Can Language Models Learn from Context Skillfully? Paper • 2604.27660 • Published 23 days ago • 163
LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model Paper • 2604.20796 • Published Apr 22 • 242
SkillClaw: Let Skills Evolve Collectively with Agentic Evolver Paper • 2604.08377 • Published Apr 9 • 291
GrandCode: Achieving Grandmaster Level in Competitive Programming via Agentic Reinforcement Learning Paper • 2604.02721 • Published Apr 3 • 630